Reasons for Restriction: Impact of expanded diagnostic ...

Received: 13 July 2018 Revised: 9 November 2018 DOI: 10.1002/eat.22988

Accepted: 13 November 2018

ORIGINAL ARTICLE

Impact of expanded diagnostic criteria for avoidant/restrictive food intake disorder on clinical comparisons with anorexia nervosa

Kendra R. Becker PhD1,2 | Ani C. Keshishian BA1 | Rachel E. Liebman PhD1,2 | Kathryn A. Coniglio BA3 | Shirley B. Wang BA1,4 | Debra L. Franko PhD1,5 | Kamryn T. Eddy PhD1,2 | Jennifer J. Thomas PhD1,2

1Eating Disorders Clinical and Research Program, Massachusetts General Hospital 2Department of Psychiatry, Harvard Medical School, Boston, Massachusetts 3Department of Psychology, Rutgers University, New Brunswick, New Jersey 4Department of Psychology, Harvard University, Boston, Massachusetts 5Northeastern University, Boston, Massachusetts

Correspondence Kendra R. Becker, PhD, Eating Disorders Clinical and Research Program, Massachusetts General Hospital, 2 Longfellow Place, Suite 200, Boston, MA 02114. Email: krbecker@mgh.harvard.edu

Funding information National Institute of Mental Health of the National Institutes of Health, Grant/Award Number: F32MH111127; National Science Foundation, Grant/Award Numbers: DGE-1433187, DGE-1745303

Abstract

Objective: Avoidant/restrictive food intake disorder (ARFID) and anorexia nervosa (AN) are restrictive eating disorders. There is a proposal before the American Psychiatric Association to broaden the current DSM-5 criteria for ARFID, which currently require dietary intake that is inadequate to support energy or nutritional needs. We compared the clinical presentations of ARFID and AN in an outpatient sample to determine how a more inclusive definition of ARFID, heterogeneous for age and weight status, is distinct from AN. Methods: As part of standard care, 138 individuals with AN or ARFID completed an online assessment battery and agreed to include their responses in research. Results: Individuals with ARFID were younger, reported earlier age of onset, and had higher percent median BMI (%mBMI) than those with AN (all ps < .001). Individuals with ARFID scored lower on measures of eating pathology, depression, anxiety, and clinical impairment (all ps < .05), but did not differ from those with AN on restrictive eating (p = .52), and scored higher on food neophobia (p < .001). Discussion: Allowing psychosocial impairment to be sufficient for an ARFID diagnosis resulted in a clinical picture of ARFID such that %mBMI was higher (and in the normal range) compared with AN. Differences in gender distribution, age, and age of onset remained consistent with previous research. Both groups reported similar levels of dietary restriction, although ARFID can be distinguished by relatively higher levels of food neophobia. Currently available measures of eating pathology may capture certain ARFID symptoms, but highlight the need for measures of impairment relative to ARFID.

KEYWORDS avoidant/restrictive food intake disorder, anorexia nervosa, outpatient, restriction, food neophobia, impairment

1 | INTRODUCTION

Avoidant/restrictive food intake disorder (ARFID) and anorexia nervosa (AN) represent the two primary restrictive eating disorders described in the 5th edition of the Diagnostic and Statistical Manual

Co-senior authors.

(DSM-5; American Psychiatric Association, 2013). Both disorders are characterized by insufficient dietary intake, and individuals with either of these disorders are at risk for serious medical sequelae including bradycardia (Cooney, Lieberman, Guimond, & Katzman, 2017), lowweight status (Norris et al., 2014), amenorrhea (Thomas et al., 2017), gastrointestinal (GI) pain/dysfunction (Norris et al., 2014), and anemia (Kelly, Shank, Bakalar, & Tanofsky-Kraff, 2014) as well as similar

230

? 2018 Wiley Periodicals, Inc.

journal/eat

Int J Eat Disord. 2019;52:230?238.

BECKER ET AL.

231

psychological comorbidities including anxiety disorders and depression (Norris et al., 2014). However, explanatory mechanisms underlying dietary restriction are hypothesized to differentiate these two diagnoses. Specifically, restriction and food avoidance in the context of ARFID is not driven by the weight and shape concerns that typify AN (American Psychiatric Association, 2013; Nicely, Lane-Loney, Masciulli, Hollenbeak, & Ornstein, 2014; Thomas, Lawson, et al., 2017). Instead, motivations for dietary restriction in ARFID are derived from research on feeding disorders.

Prior to the release of DSM-5, there was no unifying classification system for the variety of feeding disturbances observed in children. Rather, several classification systems were proposed to account for clinical presentations that were not well-addressed by the DSM-IV "Feeding Disorder of Infancy and Early Childhood" diagnosis (e.g., children with faltering growth but at normal weights, children with limited diets at normal or higher weights, individuals with feeding difficulties emerging after early childhood; for a review see BryantWaugh, Markham, Kreipe, & Walsh, 2010). Bryant-Waugh et al. (2010) suggested that three general classifications best captured the heterogeneity of feeding difficulties, and these three are now used for the example presentations of ARFID described in DSM-5: inadequate food intake (i.e., lack of interest in eating); restricted range of foods due to smells, tastes, textures, temperatures, and appearances of foods (i.e., sensory sensitivity); and food avoidance following the development of a specific eating fear (i.e., a fear of aversive consequences from eating; American Psychiatric Association, 2013).

Research on childhood and infant feeding disturbances highlighted the heterogeneity of avoidant/restrictive eating and provided a framework for the ARFID diagnosis, allowing the characterization of these eating patterns beyond childhood. However, there remains ambiguity as to how ARFID, as described in DSM-5, can be diagnosed. The current wording of the text for criterion A states that ARFID can be diagnosed if an individual is failing to meet nutritional and/or energy needs as manifested by: "significant weight loss or failure to achieve expected weight gain or faltering growth" (A1); "significant nutritional deficiency" (A2); or "dependence on enteral feeding or oral nutritional supplements" (A3). However, a "marked interference with psychosocial functioning" is listed as the final sub-criterion (A4), even though this is not a manifestation of unmet energy or nutritional needs. As such, even expert clinicians and researchers have interpreted criterion A differently, with some interpreting psychosocial impairment as sufficient for the diagnosis, and some not (Eddy et al., 2018). To reduce confusion and improve diagnostic clarity, the field is considering a proposal to expand the ARFID diagnosis to include those who, due to their eating habits, describe significant psychosocial impairment, in the absence of weight loss, nutrition deficiency, or supplement dependence (American Psychiatric Association, 2018). It is, therefore, very timely to explore how a more inclusive definition may impact the clinical presentation of ARFID.

In addition to describing and classifying various feeding difficulties, past research also compared certain presentations of feeding disorders to AN to better understand various forms of restrictive eating. A well-described presentation of inadequate food intake, Food Avoidance Emotional Disorder (FAED), could be challenging to differentially

diagnose from AN and was characterized by insufficient dietary intake in response to negative emotional states such as sadness and anxiety (Bryant-Waugh et al., 2010; Higgs, Goodyer, & Birch, 1989). Compared with those with AN, those with FAED did not report weight and shape concerns and seemed to experience higher levels of anxiety unrelated to food (Higgs et al., 1989). Children described as selective, perseverative, and/or food neophobic engaged in food refusal of nonpreferred foods and sometimes also exhibited intolerance of eating around others, excessively slow eating, obsessive and compulsive symptoms, and social difficulties (Bryant-Waugh et al., 2010), all of which could also be present in and potentially challenging to distinguish from AN (e.g., Garner & Garfinkel, 1979; Kaye et al., 2004). Finally, children with a specific fears of eating (e.g., choking, swallowing, vomiting, gastrointestinal distress) tended to present as acutely ill, having rapidly lost a substantial proportion of their body weight such that their medical presentation was similar to those with AN (BryantWaugh et al., 2010).

Since the inclusion of ARFID in DSM-5, research has continued to explore avoidant and restrictive eating in relation to AN. Generally, previous studies have reported that those with ARFID present for treatment at a younger age than those with AN (Bryson, Scipioni, Essayli, Mahoney, & Ornstein, 2018; Cooney et al., 2017; Fisher et al., 2014; Forman et al., 2014; Nakai, Nin, Noma, Teramukai, & Wonderlich, 2016; Nicely et al., 2014; Norris et al., 2014; Ornstein, Nicely, Lane-Loney, Masciulli, & Hollenbeak, 2013) and, in outpatient settings, have a longer duration of illness than patients with AN (Fisher et al., 2014; Forman et al., 2014). A recent latent class analysis of children between the ages of 5 and 13 presenting to pediatric clinics or general psychological clinics reported that restrictive eating could be separated into two distinct classes: (1) a class similar to AN characterized by body dissatisfaction, fear of gaining weight, and over exercising; and (2) a class similar to ARFID characterized by somatic concerns and low levels of weight and shape concerns (Pinhas et al., 2017). Consistent across three samples from English-speaking countries (United Kingdom, Australia, and Canada), the class similar to ARFID was younger with elevated levels of anxiety and, though not statistically significant, a longer duration of illness (Pinhas et al., 2017). Interestingly, at higher levels of care, child and adolescent patients with ARFID have often shown similar illness duration to patients with AN (Nicely et al., 2014; Ornstein, Essayli, Nicely, Masciulli, & Lane-Loney, 2017; Strandjord, Sieke, Richmond, & Rome, 2015). In contrast, in one study, adult patients with AN presenting for outpatient care reported much longer illness duration than adult patients with ARFID (Nakai et al., 2016), suggesting that age of presentation and treatment setting may impact comparisons between AN and ARFID.

Similarly, some studies suggest that the proportion of males versus females in ARFID is higher than the proportion of males versus female in AN, particularly in outpatient samples (Bryson et al., 2018; Fisher et al., 2014; Forman et al., 2014; Nicely et al., 2014; Norris et al., 2014; Ornstein, Nicely, et al., 2013). However, the gender difference between ARFID and AN is less pronounced in adult samples (Nakai et al., 2016) and in patients requiring acute medical hospitalization (Strandjord et al., 2015).

All currently published studies have compared underweight individuals with ARFID to those with AN. Importantly, unlike AN, ARFID

232

BECKER ET AL.

can be diagnosed across the weight spectrum because low weight or faltering growth is only one of the potential expressions of the diagnostic criteria. Alternative manifestations defined by the current DSM-5 criteria include nutritional deficiency and reliance on nutritional supplements or enteral feeding. For example, dietary avoidance of all but preferred foods--typically highly processed snack foods--can be associated with nutritional deficiencies in those with overweight or obese presentations of ARFID (Thomas & Eddy, 2019). Further, extending criterion A to include psychosocial impairment would allow individuals who are meeting energy and nutritional needs to be diagnosed with ARFID. As suggested in the proposal to revise DSM-5 criteria, the following would represent significant impairment for diagnosis: "Inability to participate in normal social activities, such as eating with others, attending school or work or sustaining relationships as a result of the eating disturbance would indicate marked interference with psychosocial functioning. Substantial disruption of family functioning, such as marked restriction of foods permitted in the home or inordinate accommodations to provide foods from specific grocery stores or restaurants, may also satisfy criterion A4 (American Psychiatric Association, 2018)." Research examining how those with a broader range of avoidant and/or restrictive eating and body weights may compare to those with AN is needed to more adequately represent the heterogeneous presentations of ARFID.

Psychometric analyses have revealed that, consistent with diagnostic criteria, individuals with ARFID report less body image disturbance, drive for thinness, and concern about weight/shape and eating compared with those with AN, as well as fewer bulimic behaviors (Nakai et al., 2016; Ornstein, Nicely, et al., 2013). However, because ARFID is a newly defined disorder, retrospective study samples relied on clinical evaluations and measures that were designed for the assessment of psychopathology more reflective of AN than ARFID. Therefore, more research is necessary to describe ARFID in children, adolescents, and adults diagnosed through nonretrospective evaluations and with measures that may also assess ARFID symptomatology.

Although mood and anxiety disorders are highly comorbid with AN (Blinder, Cumella, & Sanathara, 2006; Hudson, Hiripi, Pope, & Kessler, 2007), less is known about comorbid internalizing disorders in relation to ARFID. Some research suggests that adolescents and children with ARFID are more likely to have anxiety disorders and less likely to have mood disorders compared with patients with AN (Fisher et al., 2014; Nicely et al., 2014). Similarly, the few studies that have explored self-reported mood and anxiety symptoms in those with ARFID have found that individuals with ARFID may be less likely to endorse depressive symptoms than anxiety symptoms (Cooney et al., 2017; Nicely et al., 2014). There are currently no self-report data on anxiety and depressive symptoms in adults with ARFID; however, there is some data on the relationship between ARFID symptoms and mood disturbances in adult community samples (Zickgraf, Franklin, & Rozin, 2016). Adults who self-identified as picky eaters and endorsed at least one consequence of their picky eating (e.g., weight loss, nutritional deficiency, reliance on nutritional supplements, or psychosocial impairment in work or with family/friends) endorsed comparable levels of internalizing distress (i.e., level of negative emotions experienced) as adults who reported symptoms consistent with traditional eating disorders (Zickgraf et al., 2016). Additional research is needed

to determine if observed differences in anxiety and depression symptoms between individual diagnosed with ARFID or AN are evident across ages and eating-disorder severity to better explain clinical impairment associated with ARFID and aid in differential diagnosis.

The current literature exploring similarities and differences between individuals with ARFID or AN has revealed important information about the clinical picture of these two restrictive eating disorders. However, most existing study designs rely on retrospective chart reviews of patients who presented prior to the inclusion of ARFID in DSM-5. It is also largely unknown if results from previous studies apply only to the low-weight and/or the child/adolescent ARFID presentations, or whether these findings also generalize to normal/overweight and adult presentations. Importantly, these limitations are partially related to the remaining ambiguity around psychosocial impairment as a sufficient criterion for an ARFID diagnosis.

It is prudent to fill the gaps in our understanding of ARFID because there is a current proposal to alter the DSM-5 ARFID diagnosis by removing the phase in the stem of criterion A which stipulates that the eating or feeding disturbance "manifests as a persistent failure to meet appropriate nutritional and/or energy needs" (American Psychiatric Association, 2018). In support of this proposal, prior classification schemes for childhood feeding disorders clearly identified children with selective eating and food neophobia who were not low weight or nutritionally deficient (Bryant-Waugh et al., 2010) and recent research suggests that selective eating in adults can be associated with significant psychosocial impairment (Zickgraf et al., 2016). It is, therefore, imperative to explore how a more inclusive definition of ARFID compares to AN to help inform how the proposed revision may impact the presentation of and impairment from restrictive eating disorders. Thus, the aim of the current study was to further differentiate restrictive eating disorders by comparing the clinical presentations of children, adolescents, and adults with either ARFID or AN seeking treatment at an outpatient eating-disorder clinic.

In the current study, we compared participant responses to selfreport measures of eating, mood, anxiety, and clinical impairment. Unlike previous studies that have used relatively homogenous samples, we included participants across developmental and weight spectrums. To explore how the proposal to include psychosocial impairment as a criterion sufficient for an ARFID diagnosis may impact group comparisons on age of presentation, gender distribution, body weight, levels of anxiety, depression, and eating pathology, we assigned an ARFID diagnosis when nutritional and/or energy needs were unmet as well as when marked interference with psychosocial functioning was evident despite adequate energy and nutritional intake. We expected that this broadening of the diagnostic criteria would result in higher body weights, on average, for those diagnosed with ARFID compared with AN. As in previous studies, we hypothesized that, relative to individuals with AN, those with ARFID would be younger at treatment presentation and would report an earlier age of illness onset. We also hypothesized that patients with ARFID would score lower across measures of traditional eating pathology, higher on a measure of food neophobia, and, consistent with previous research, would report more anxious symptoms but fewer depressive symptoms than individuals with AN.

BECKER ET AL.

233

2 | METHODS

2.1 | Participants

Participants were individuals or parents who consecutively called the intake line of an outpatient clinic in a tertiary care hospital and were scheduled for an evaluation appointment with a clinician. We did not include individuals who were evaluated but diagnosed with an eating disorder other than AN or ARFID (e.g., bulimia nervosa, binge eating disorder), as this was outside the scope of this article. As part of standard clinical care, all participants received a link via e-mail prompting them to complete an online battery of self-report questionnaires prior to their evaluation appointment. Upon following the link in the e-mail, participants were provided a description of the data repository including IRBapproved consent/assent documents and asked to indicate their consent via checking a box indicating if they did or did not want their responses to the questionnaires to be used for research purposes. Participants were encouraged to complete the questionnaires before being evaluated, because responses would be used for diagnosis and treatment planning. All individuals who provided consent for their responses to be used for research were included in this study, even if they were not followed by a clinician after their evaluation appointment.

Between 2014 and 2017, 138 individuals with either AN or ARFID agreed to have their responses on these questionnaires used for research and 25 did not. Independent t-tests revealed that these patients did not differ on study measures compared with those who were counted as participants in this study (all p's > .42). Based on a formal clinical evaluation by a PhD- or MD-level clinician, 67 (49.46%) individuals were diagnosed with ARFID as their primary DSM-5 diagnosis and 71 (51.54%) individuals were diagnosed with AN. Clinicians used an evaluation template derived from DSM-5 criteria to diagnosis ARFID or AN. Questions related to the diagnosis of ARFID included: number of foods eaten regularly from each of the five major food groups (i.e., fruits, vegetables, proteins, grains, and dairy); self-reported sensitivity to the appearance, taste, texture, and smell of foods; appetite and enjoyment of food; experience of food-related trauma; presence/absence of nutritional deficiencies; reliance on nutritional supplements; weight status and weight history; and psychosocial impairment including, but not limited to, avoidance of social events/ holidays for fear of being around new foods, frequency of skipped or forgotten meals, and arguments around meal times or accommodations to food/eating preferences and fears. Questions used to help diagnose AN included weight history and status, fear of weight gain, body image concerns, appearance comparisons, body checking and avoidance behaviors, typical daily intake, dietary rules, exercise behaviors, binge eating, and compensatory behaviors. For the most accurate diagnosis, self-reported nutritional deficiencies and growth trajectories were checked against recent medical visits when available. The Partners Human Research Committee approved this data collection.

2.2 | Measures

Most measures were included in the online battery of self-report questionnaires beginning in 2014. However, the Food Neophobia Scale (FNS) and the Center for Epidemiological Studies-Depression Scale Revised

(CES-D) were added in 2017. Thus, fewer participants completed measures added later in data collection. All participants in the study completed the below measures and we asked that younger children complete the measures with assistance from their parents, if needed.

We collected self-reported age, gender, race, height, weight, and age of illness onset on a demographics questionnaire. Age of illness onset was assessed with the question: "How old were you when your eating or feeding disorder first started?"

2.2.1 | Eating pathology measures

The Eating Disorder Examination-Questionnaire (EDE-Q 6.0; Fairburn & Beglin, 2008) is 28-item measure that assesses the frequency and severity of eating-disorder symptoms on four subscales (Restraint, Eating Concern, Shape Concern, and Weight Concern) as well as a Global score (in the current sample, internal consistency for all scales was greater than .84). The Eating Pathology Symptoms Inventory (EPSI; Forbush et al., 2013) is a 45-item measure that examines eatingdisorder pathology using eight scales: Body Dissatisfaction, Binge Eating, Cognitive Restraint, Purging, Restricting, Excessive Exercise, Negative Attitudes toward Obesity, and Muscle Building (internal consistency values in the current sample were above .85 for all subscales except Purging and Muscle Building where = .67 and = .60, respectively). Previous studies have used these measures in young adolescents and children, suggesting that participants as young as 11 on the EDE-Q (Wang & Borders, 2018) and 10 on the EPSI (Coniglio & Becker et al., 2018) are able to report on their symptoms.

2.2.2 | Acceptability of novel foods and appetite for palatable foods

The Food Neophobia Scale (FNS; Pliner & Hobden, 1992) is a 10-item measure that assesses willingness to try unfamiliar foods (in the current sample, = .93). The Power of Food Scale (PFS; Lowe et al., 2009) is a 15-item measure that assesses food responsiveness using three subscales: Food Available (e.g. "I find myself thinking about food even when I am not physically hungry"), Food Present (e.g. "I get more pleasure from eating than I do from almost anything else."), Food Tasted (e.g. "It's scary to think of the power that food has over me."; in the current sample, = .87, = .83, = .85, respectively) and a total score (in the current sample, = .92). Children as young as seven have completed the FNS with help from their parents (Koivisto & Sj?d?n, 1996). Children and adolescents between the ages of 11 and 18 have completed the PFS, and their results suggested the same 3-factor structure found with adult participants (Mitchell, Cushing, & Amaro, 2016).

2.2.3 | Anxiety, depression, and clinical impairment measures

The State?Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983) is a 21-item measure of anxiety symptoms (in the current sample, = .93). The Center for Epidemiological Studies Depression Scale Revised (CES-D; Radloff, 1977) is a 20-item measure of depressive symptoms (in the current sample, = .92). The Clinical Impairment Assessment (CIA; Bohn & Fairburn, 2008) is a 16-item measure of psychosocial impairment due to eating-disorder symptoms (in the current sample, = .95). The CES-D has been

234

BECKER ET AL.

selected for use in studies with participants as young as 9 years of age and has demonstrated good reliability, internal consistency, and validity in predicting major depressive disorder in youth (Compas et al., 2015). The STAI has been used with children as young as 13 (Martin, Viljoen, Kidd, & Seedat, 2014) and the CIA has been used in adolescents as young as 15 (Reas, Stedal, Lindvall Dahlgren, & R?, 2016). We asked that younger participants completed with the assistance of their parents, though we do not have data on which families did so.

2.3 | Data analysis

We conducted all analyses in SPSS version 23 (IBM Corp, 2016) and R (R Core Team, 2018) using the "effsize" package (Torchiano, 2016). We analyzed group differences with independent Student's t-tests and Analyses of Covariance (ANCOVAs) with Bonferroni correction to protect against Type I error. Based on previous research suggesting that individuals with ARFID tend to be younger at treatment presentation than individuals with AN (Fisher et al., 2014; Nicely et al., 2014; Norris et al., 2014; Ornstein et al., 2017), and because our sample was heterogeneous with regard to weight status, we entered age and percent median BMI for age (%mBMI) as covariates. We calculated %mBMI using patients' self-reported age, height, and weight using the following formula: BMI (kg/m2)/median BMI for age (as predetermined by the National Center for Health Statistics) ? 100. Because the growth charts available from the Center for Disease Control and Prevention (CDC) extend to just 20 years of age, we used 20 as the reference age to calculate the %mBMI for any participant 20 years old or older.

3 | RESULTS

Mean age of the sample at treatment presentation was 22.22 (SD = 11.88, range 10?78) and 73.8% (n = 96) were female. The ARFID group presented for treatment at a significantly younger age compared with the AN sample (AN = 26.38, ARFID = 18.01; p < .001; d = 0.72). Of the 67 individuals with ARFID, 34 (50.8%) were female and of the 71 individuals with AN, 67 (94.4%) were female. The ARFID sample included a significantly higher proportion of males compared with the AN sample (p < .001; = 0.51). Over 90% of both groups (AN = 92.5%, ARFID = 93.0%) identified as Caucasian.

Age of eating-disorder onset was significantly younger in individuals with ARFID compared with individuals with AN (AN = 16.38 years, ARFID = 8.30 years; p ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download